What should companies consider before investing in a BI solution?

Before buying anything, decision makers should inventory everything they already have for its actual or potential role in delivering more intelligence to the business. Things as basic as tracking the use of shared files by user or department can tell a business a lot, as can tracking who accesses what information at the company Web site. But many companies have not even started to do such relatively simple tasks in a consistent well-managed way.

In addition, businesses serious about maximizing the business value and ROI of any investment in BI solutions should work with resellers, integrators their own IT teams and their peers to address issues of data quality, consistency and security, proactively and effectively. If there is an area of business or technology beyond computer programming in which the admonition about "garbage in, garbage out" matters, it's BI.

First thing to ask yourself - how much do you want / can you spend? It seems complete BI solutions come in roughly two sizes - $50K +/- for Open Source-ish (Pentaho, Jaspersoft, etc.) and MSFT based solutions and $250K+ from the big boys (Oracle, SAP, Microstrategy, etc.)

Most of the major players have very similar functionality for basic reporting and analysis, but they do differ significantly in how they will fit within your environment. For instance, the high end solutions tend to require proprietary DBs or utilize commercial ones like Oracle or DB2, obviously the MSFT solutions are based around SqlServer and the Open Source solutions work with things like MySQL. You'll also want to figure out what sort of out of the box connections each of the platforms has since that can help speed implementation if you have a lot of systems you are trying to integrate.

Lastly you should consider what, if any, advanced functionality you need. Things like data mining and advanced statistical analysis. This is a place where vendor do differentiate themselves

The first thing is to define your requirements. You should have a very clear idea of what you want before parting with a single penny. In the case of BI, you need to understand what information you are looking to get out - is it financial, production related, sales performance etc. (Of course it may be many of these)

Once you have the main thrust of what you hope to get, (lets say sales), you then need to establish in more detail specifically what you are looking to obtain from the system e.g. sales staff perfomance via various metrics, effectiveness of marketing, customer retention / satisfaction etc.

It may sound bizarre, but most companies actually really don't know what they want from a system when they buy it, and are astonished when it doesn't deliver anything of value despite having spent vast sums of money on the project.

They only start to get useful information from the system some time later, at which point they then realise what else they could have included - but then all too often, they are restricted in what they can achieve due to limitations caused by not knowing in more detail from the start, what they really wanted.

By the way, Focus Expert and BI Maven Wayne Kernochan just posted a great brief on getting good data out of your BI systems -- it's at http://focus.com/c/Bwo/. Enjoy!

1) Do we have the patience to wait for the results.Its not a magic wand which will make the business grow in a fortnight.
2) should we go for inhouse BI developers or outsource. If we want inhouse, do we have the setup/infrastructure.
3) with the changes in business trend, how do my BI solution adapt to the changes and what is the price I have to pay?
4) What is my expectation from the solution.Is my problem statement defined?
5) How much value addition will the solution make in the long run and what would be my ROI and by when.

Finally all BI solutions will give you some interesting facts. It is up to you to take the right decission. So your decission makes or breaks the reputation of the solution provider.

I highly recommend viewing BI SOLUTION as inclusive of people, and process - not just technology.

From that perspective, I would ensure the following Business (people/business process) tenants:
* Scalability
* Repeatability
* Auditability
* Fexibility
* Consistency

From a technical perspective I would ensure the following tenants:
* High Availability
* Fault-Tolerance
* Scalability
* Auditability
* Flexibility

It's a matter of asking the question: What happens to the BI Solution when the business changes their mind? Does it cause re-engineering? Do you have to hire new people? Does it cause massive infrastructure changes?

These are the types of questions that both vendors and consulting companies typically skirt around. Understanding these answers is vital to the proper SOLUTION.

I've seen too many BI Solutions "been dropped completely" only to be restarted from scratch because someone up-front didn't consider the above tenants.

I would suggest the study of MPP, VLDW, SEI/CMMI, Six Sigma, TQM, PMP, Lean Initiatives, Cycle Time Reduction, Process Management and Improvement, and Agile Methodologies before truly embarking on an enterprise wide solution.

Cheers,
Dan Linstedt
PS: You can follow my BI based business blog at: http://empoweredHoldings.com
and my Data Vault Modeling blog at: http://danLinstedt.com

Hi,

Every decision that involves important tecnological and strategic effort should be driven by a good diagnose from a particular need. If it's a BI solution or just simple process improvement doens't matter. What's critical is that the Entreprise Decision Making Process identified, in this case a need. After the need is identified there are solutions for every type of problem, nowadays.

Thanks for the question.
Best regards,
Ricardo Andorinho

For BI to be effective and efficient, all data systems, relevant to the outcome, must be integrated and interopperable. This is necessary because any disparate or silo data cannot support decisions if it is not connected. So the prerequisite is that all systems be integrated before BI is selected or placed in service.

Understand who the consumer of the data will be and what they need. Understand what is the value of having the right data in the right format for those data consumers. What is the impact to the company? Is it more sales? If so, how much more? Is it better customer service? How will this impact the company's bottom line?

Understand what the value is for having BI. Set goals and time-lines that are reasonable and allow for you to match the budget (value) with the cost of implementation.